Pricing

from $2.80 / 1,000 version returneds

Try for free

Go to Apify Store

Changelog to JSON - Release Notes, GitHub Releases, npm & PyPI

Try for free

For AI agents and developers upgrading a dependency: give a changelog page, GitHub repo or npm/PyPI package, get one JSON row per version with date and added / fixed / removed / breaking items, only versions newer than yours if you ask. On 41 real changelogs, 371 of 375 versions returned were real.

Pricing

from $2.80 / 1,000 version returneds

Rating

0.0

(0)

Developer

NeverEmpty

Actor stats

Bookmarked

Total users

Monthly active users

17 days ago

Last modified

Categories

Developer tools

Automation

Changelog page URLs

urls

Optional

Changelog or release-notes pages, one per line: a CHANGELOG.md / HISTORY.md / CHANGES.rst file (a github.com/.../blob/... link is read as the raw file), a /changelog or /release-notes web page, or a single release-notes page. A github.com/owner/name link is read as a repository (see Repositories); an npmjs.com or pypi.org package link is read as that package. If no input field is filled at all, the example https://github.com/vitejs/vite/blob/main/packages/vite/CHANGELOG.md is used (newest 3 versions).

Type:array

GitHub repositories

repos

Optional

GitHub repositories as owner/name (for example facebook/react), one per line. Both the GitHub Releases list (official API) and the changelog file in the repository (CHANGELOG.md, CHANGES.md, HISTORY.md, NEWS.md, RELEASES.md, CHANGELOG.rst and similar) are read and matched by version; the source column says where each row came from. For a package inside a monorepo write owner/name/path/to/package to look for the changelog file there first.

Type:array

npm packages

npmPackages

Optional

npm package names (for example react or @anthropic-ai/sdk). The package's repository is read from the npm registry and then handled like a GitHub repository (the changelog in the package's own folder first). The latest published version is added to every row as registryLatestVersion.

Type:array

PyPI packages

pypiPackages

Optional

PyPI project names (for example django). PyPI's public release feed gives the newest 40 version numbers and upload times only, without release notes; give the project's GitHub repository in Repositories or its changelog URL in Changelog page URLs to get the notes.

Type:array

Only versions newer than

sinceVersion

Optional

Return only versions greater than this one (for example 1.4.0 or v2.3.1), compared as version numbers (1.10.0 is newer than 1.9.0; a pre-release such as 2.0.0-rc.1 is older than 2.0.0). Rows without a version number are left out when this is set. A date-style API version such as 2024-09-30.acacia is compared by its date.

Type:string

Only versions released after

sinceDate

Optional

Return only versions dated after this day (YYYY-MM-DD). Rows without a date are left out when this is set.

Type:string

Max versions per input

maxVersions

Optional

At most this many versions per changelog, newest first (1 to 1,000). Each returned version is one charged row.

Type:integer

Minimum:1

Maximum:1000

Default:20

Include pre-releases

includePrereleases

Optional

Also return versions marked as pre-releases (alpha, beta, rc, canary, next, dev, nightly, or a GitHub release marked pre-release). Off by default.

Type:boolean

Default:false

Include the Unreleased section

includeUnreleased

Optional

Also return an "Unreleased" / "Upcoming" section at the top of a changelog (isUnreleased is true on that row). Off by default.

Type:boolean

Default:false

Include the full notes text

includeBody

Optional

Return the whole text of each version as Markdown in body (up to 20,000 characters). Turn off for smaller rows with only the sorted lists (added, fixed, breaking and the others).

Type:boolean

Default:true

Return the whole text when it cannot be split into versions

returnWholeTextIfUnsplit

Optional

When a page has no heading with a version number or a date, return its whole text as one row (status unsplit, charged as one row) instead of a free row that says it could not be split.

Type:boolean

Default:false

Monitor mode: only versions not seen before

onlyNewVersions

Optional

Return only versions that did not appear in the previous runs with the same watch name. The first run returns the newest versions (up to Max versions per input) and remembers every version on the page. A run with no new version returns a free row saying so and charges only the run start fee.

Type:boolean

Default:false

Watch name

watchName

Optional

Name of the remembered state used to compare runs (letters, digits, dot, dash, underscore; up to 40). Setting it (or turning on monitor mode) fills changeType (first-check or new). Use a different name for each list you check on its own schedule. With monitor mode on and no name, the name "default" is used.

Type:string

Forget what this watch remembered

resetMonitoringState

Optional

Start this watch over: forget the remembered versions before this run.

Type:boolean

Default:false

GitHub token (optional)

githubToken

Optional

Optional. Without a token, GitHub allows about 60 API calls per hour for the IP address the run uses; this Actor checks the remaining allowance first (the check itself is free) and, when it is used up, reads only the changelog files and says so. A personal access token with no scopes raises the allowance to 5,000 per hour. It is sent only to api.github.com.

Type:string

Inputs read at the same time

maxConcurrency

Optional

How many inputs are read in parallel (1 to 8).

Type:integer

Minimum:1

Maximum:8

Default:4

Time limit per request (seconds)

requestTimeoutSecs

Optional

How long to wait for one answer (5 to 60 seconds). HTTP 429, 500, 502, 503 and 504 are asked again up to two more times.

Type:integer

Minimum:5

Maximum:60

Default:25

TOON to JSON (converter)

room_js/toon-to-json-converter

This actor allows you to convert TOON format to JSON

JavaScript Room

Weather Forecast Scraper - NWS Weather Forecast API, Hourly

neverempty/us-weather-forecast-api

For logistics, field-service scheduling, energy demand and event planning: 7-day and hour-by-hour forecasts for any US coordinates from the National Weather Service's own API. Fahrenheit and Celsius in every row. US government public domain: no API key, no rate-limit contract, no proxies.

NeverEmpty

GitHub Release Notes & Changelog Scraper

happyfhantum/github-changelogs

Track GitHub releases and changelogs to spot product updates, launches, and breaking changes.

Kelsey Todd

GitHub Release Notes & Changelog Scraper

w3crawler/github-changelogs-scraper

w3crawler

ESPN Injury Report Scraper & Change Monitor - NFL NBA MLB NHL

neverempty/espn-injury-report-scraper

For fantasy tools, betting models and news desks: every current ESPN injury report across six leagues in one row shape - 1,261 reports on 2026-09-06 (NFL 800, MLB 283, NBA 75, NHL 61, WNBA 39), each with the player ESPN id. Monitoring returns only what changed, so you never pay twice. No API key.

NeverEmpty

npm & PyPI Package Monitor — Versions & Downloads

vertaizen/package-registry-monitor

Monitor npm & PyPI packages via the open registry APIs — latest version, publish date, monthly downloads, dependencies & maintainers. Monitor mode returns only CHANGES: new versions & download swings. Supply-chain security, devtools & investor signals. Pay per package.

Diego Moragues

Changelog / Docs Monitor API

s3nafps/changelog-docs-monitor-api

Extract recent product, docs, changelog, RSS, and GitHub release updates from public sources.

mohamed senator

Service Status Checker - Is It Down? Status Page & Outage API

neverempty/service-status-checker

For AI agents: when a call to GitHub, OpenAI, AWS, Stripe or 400 other services fails, check what its official status page says now - operational, degraded, outage or maintenance, with active incidents and affected components. Monitoring mode returns only changes.

NeverEmpty

HTML Table to JSON & CSV - Web Table Extractor

neverempty/html-table-to-json

For AI agents and data pipelines: give a URL, get every HTML table on the page as JSON rows keyed by column name, with merged cells expanded, multi-row headers joined, numbers typed and footnotes removed. On 48 real pages, 454 tables matched what a browser draws, cell by cell. CSV too. Monitor mode.

NeverEmpty

YouTube Transcript Scraper API - Captions and Subtitles

neverempty/youtube-transcript-reliable

For RAG pipelines, LLM training and content repurposing: transcripts of public YouTube videos - manual and auto-generated captions - as clean text or timestamped lines. When one fails you are told which failure it was: private video, no captions, or blocked. 14 of 15 gave 83,647 characters in 48.7s.

NeverEmpty